Why energy storage is a dead-end industry

A report by IFK Berlin on the energy return on investment (EROI) of battery energy storage, when used to balance intermittent renewable energy on a grid scale, suggests it may not be viable.

An IFK Berlin report on the energy return on investment (EROI) of battery energy storage suggests it isn’t viable when used to balance intermittent renewable energy. Photo credit: Guto

Could energy storage send us back to the Stone Age? Galling as it may seem to those of us who view storage as the solution to the problem of renewable energy intermittency, and hence the key to a carbon-free future, there is a growing body of evidence that suggests this might indeed be the case.

The studies have nothing to do with energy storage’s main preoccupation at the moment, which is cost. Instead, they deal with a much more fundamental issue: how much energy it takes to be able to store the energy in the first place.

The concept of energy return on investment (EROI, also called energy returned on energy invested) is critical to energy storage because it provides a measure of whether a particular technology might be appropriate for use at scale.

In essence, if it takes more energy to create a given storage mechanism than the mechanism could ever deliver over the course of its life, then the net result of using the technology is that it will cannibalise power rather than return it to the system.

Energy vampires

Such energy vampires may be appropriate for specific point applications but are clearly unsustainable on a grand scale. As volumes grow, more energy and more energy will be tied up in creating the storage technology and less will be left to do useful work.

EROI measures are not just relevant to energy storage, but to power generation technologies in general.

In the past, however, EROIs do not appear to have been studied exhaustively because the values attributable to traditional energy plants are relatively good.

Put another way: the amount of energy it takes to build a coal, diesel, gas, hydro or nuclear power plant will be repaid many times over during the lifetime of the facility.

The picture is changing with the advent of renewable energy technologies that require substantial energy inputs for their manufacture.

The energy used to produce energy

Although, unsurprisingly, working out exactly how much energy goes into the making and operation of a power plant is far from easy. Studies of the subject are frequently bogged down with discussions of what measures should or should not be included.

For example, should the EROI of a coal-fired plant include the energy needed to build the truck that hauls coal from the mine? Despite these difficulties, a number of recent studies have tried to put definitive EROI numbers on different generation technologies.

What seems generally the case, and is intuitively obvious, is that EROIs are higher for plants built with a high degree of mechanical engineering as opposed to manufacturing, and that function for a relatively long time on an energy-rich power source.

Thus, a 2010 study offered up an EROI of 100 for hydro power, 80 for coal and between 50 and 75 for nuclear in the USA. This means a hydro plant would deliver 100 times more energy than it took to build, and so on.

Such high EROIs are hardly surprising for these power sources. A dam might need a fair amount of engineering muscle to build in the first place, but thereafter the energy is essentially free and the plant could last 100 years or more.

Energy return on investment and renewable energy

With newer renewable technologies, though, EROIs tend to drop sharply. Wind turbines and solar panels require energy-intensive materials processing and precision manufacturing, and currently are built with an expected lifespan of around 25 years.

In the 2010 study, this translated into EROIs of 18 for wind power and 6.8 for PV. That’s not so good. But there is still a return. You still get more energy out than you put in. So what is the problem?

The first snag is that, as we know, it is practically impossible to rely too heavily on wind or solar PV without some form of storage in place to cope with the intermittent nature of generation. And many types of storage, it turns out, also have poor EROIs.

As previously reported in Energy Storage Reportresearchers at Stanford University last year calculated the energy stored on investment (‘ESOI’) for pumped hydro and five ‘promising battery technologies’.

While pumped hydro came out with a healthy ESOI of 210, the battery technologies scored between 10, for lithium-ion, and two, for lead-acid batteries, because of their tendency to wear out after a relatively limited number of cycles.

Batteries: a marginal benefit?

Given the measurement challenges mentioned above, it would be unwise to assume these values are directly comparable to the 2010 EROI figures. Nevertheless, it is clear there is only a very marginal benefit in using batteries to store energy.

Indeed, as we have covered previously, ePower Engine Systems executive vice president and long-standing electric vehicle critic John Petersen has used the Stanford research to claim volume production of battery-powered cars would be a mistake.

Meanwhile the Stanford University team last autumn revisited its figures and found that, for wind farms at least, it would be more cost effective to simply curtail generation rather than invest in the power needed to make batteries for storage.

That still leaves PV, though. And anyway, even if batteries only have marginal ESOIs, provided those values are still positive wouldn’t there still be a case for using them? Unfortunately for the battery industry, this is where things get even more complicated.

It turns out merely having a positive ESOI or EROI is not enough.

Energy needed by society

Last year a team led by Daniel Weißbach of the Institut fur Festkorper-Kernphysik in Berlin, Germany, published a paper that considered how much energy a society would need to have left over, after power production, to maintain itself.

Clearly this depends somewhat on the nature of the society, since indigenous tribes in Brazil can survive without any electricity at all.

But for high-energy-use countries such as Germany or the USA, Weißbach et al calculated that only energy sources with an EROI of seven or above would deliver enough extra power to keep things running on a business-as-usual basis.

The study actually used a measure called ‘energy money returned on invested’ (EMROI), where electrical energy inputs and outputs are weighted threefold over thermal energy.

The ‘economic threshold’ using this measure, which in general somewhat flatters low-carbon energy sources, was found to be about 16. This is already bad news for PV and biomass, which scored 5.6 and 4.8 respectively in Weißbach et al’s analysis.

Adding buffering into the equation

But then the researchers added ‘buffering’ into the equation, to account for output variability. Buffering, in this context, was taken to mean the need for overcapacity and storage to enable on-demand dispatch of the energy source.

The only storage technology considered was pumped hydro, which as we have seen is pretty efficient. Even so, the need for buffering has a devastating effect on the EMROI values for renewable energy sources.

While the values for coal, combined-cycle gas and nuclear remain unchanged, at 49, 85 and 100 respectively, wind power plunges from an un-buffered level of 42 down to 11, below the economic-usefulness threshold.

Medium-sized hydro remains more competitive than nuclear, but drops from an EMROI of 147 to a buffered value of 105.

And while concentrated solar power (CSP) also remains above the economic-usefulness threshold, falling from 51 to 25 in EMROI, the buffered figure is meaningless since it is almost impossible to envisage CSP being paired with pumped hydro storage.

This is potentially very serious stuff. If accurate, it means that the pressure-group ideal of a fully renewable, nuclear-free future is not just unachievable, but could lead modern societies into a death spiral of diminishing energy returns along the way.

Negating the use of batteries

In addition, the calculations essentially negate the use of batteries for anything other than powering your mobile phones and laptops.

At grid scale, only pumped hydro, compressed air energy storage and possibly molten salt have decent enough EROIs to be viable. Last but not least, it is important to note that these are not crackpot-authored back-of-a-napkin scribbles.

The Weißbach study, for example, used transparent calculations based on published data and was accepted for publication in Energy, a peer-reviewed journal.

And the authors were able to put up a spirited defence of the work in the face of a critique by Marco Raugei of the Faculty of Technology, Design and the Environment at Oxford Brookes University in the UK.

In a nutshell, the current work on EROI and similar measures suggests there could be something fundamentally wrong with the move to renewable generation sources in general and the use of energy storage to balance intermittent sources in particular.

The question is: what could and should the energy storage industry do about it?

Find out in our next article

Written by Jason Deign

4 thoughts on “Why energy storage is a dead-end industry

  1. Off the top of my head, design wind turbines and PV systems for easy recyclability. Virgin ores are more difficult to mine and process than, say, scrap metal or re-using old but still viable parts.

    Not the most exciting of news but an important issue to tackle no less. Hopefully engineers far and wide will see this and solve this problem.

  2. As with most mathematical analyses, it is prudent to question the assumptions being made.

    This research assumes that the energy invested will not reduce over time as new design and technology advancements occur. Most technologies have experienced this (semiconductors in particular), and ongoing research into organic PV and organic storage suggest to me that this will continue far into the future.

    Secondly, an increased lifespan of wind, pv, and storage would significantly alter the amortization – this is certainly also the case when looking at the ongoing research (eg. better anode/cathodes for lithium), and if these replace previous energy investments and are not additional, they are clear improvements.

    Third, the externalized GHG costs of fossil and hydro also have implications on the abilities of civilizations to sustain themselves. Energy focus alone is overly reductivist for the assertion being made.

    • Good points, Tom. Indeed, we’ll be covering some of the potential weaknesses of the EROI analysis in more detail next week… stay tuned!


      • Thank you for this very insightfull article and also Tom for th ecoments.

        However teh basic rules of thermodinamics seem to be still valid.

        We can not produce Energy, but only convert it and no conversion is 100 % efficient.

        In chemistry and physics one can assess the enrgy content of soemething in different states, and also conversion efficiency.

        If Hydrogen contains so uch energy you have to put that amount of enrgy into it first. and if you want to liquify it you have to use enrgy for that.

        There are studies that claculated the efficiency for generating hydrogen from electricity generated by wind generators and then converting it back to electricity when there is no wind – the efficiency was around 30%.

        One can not liquify a gas without compressing it which means cooling it which wil always lead to a loss of energy, mainly in form of heat.

        The limits of these technologies are not new the processes are pretty well known.

        The whole energy storae indutry is bogous if they do not engage in the discussion ehere these limits anr and how realistic it is to overcome these.

        If that does not happen there should be no more subsidies, IMHO.

        It would be nice to hear more about these topics anyways.

        But we have to limit our energy consumption radically, immidiately.

Let us know what you think. Please leave a comment.